Abstracts
Recently, we have reported the production of a microporous high purity silica powder from the acid leaching of glasses with average pore size around 2.5nm and specific surface area of 420m2/g (BET). The employed glasses derived from the melting of two types of waste from the industrial processing of Brazilian oil shale (retorted oil shale and the top fraction of the intermediate layer of Irati Formation). Depending on the proportion of the two wastes employed in the formulation of the glasses, either a silica gel or a powdered one is obtained, after leaching. The acid leaching of those glasses with hydrochloric acid, at 90 °C, was used to produce powdered microporous silica. In the present work it is studied the effect of aging time and temperature on the morphology and structure of the obtained powdered silica. Aging studies were performed in two different media, an acidic (hydrochloric acid) and a basic one (ammonium hydroxide) for different periods of time and temperatures. XRD, SEM/EDX, TEM, specific surface area measurements and DTA/TGA were used to characterize these materials. The results have shown a decrease of specific surface area with increasing time and temperature. Apparently, this behavior may be associated with dissolution and re-precipitation mechanisms.
glass leaching; microporous silica
Recentemente apresentamos a produção de pós microporosos de sílica de alta pureza a partir de lixiviação ácida de vidros, com tamanho médio de poros próximo de 2,5 nm e 420 m2/g (BET) de área de superfície específica. Foram usados vidros a partir de fusão de dois tipos de resíduos de processamento industrial de xisto oleígeno brasileiro (xisto oleígeno retortado e fração superior da camada intermediária da formação de Irati). Dependendo da proporção dos dois resíduos empregados na formulação dos vidros, é obtida sílica na forma de gel ou de pó, após lixiviação. A lixiviação desses vidros com ácido clorídrico a 90 oC foi usada para produzir sílica microporosa em pó. Neste trabalho é estudado o efeito do tempo e da temperatura de envelhecimento na morfologia e na estrutura da sílica obtida na forma de pó. Estudos de envelhecimento foram efetuados em dois meios diferentes, um ácido (ácido clorídrico) e um básico (hidróxido de amônio) a diferentes tempos e temperaturas. DRX, MEV/EDX, MET, medidas de área de superfície específica e ATG/ATD foram usados para caracterizar esses materiais. Os resultados mostraram uma diminuição na área de superfície específica com aumento de tempo e de temperatura. Esse comportamento pode ser aparentemente associado a mecanismos de dissolução e re-precipitação.
lixiviação de vidros; sílica microporosa
Powdered microporous glasses: changing porosity through aging
(Pós de vidros microporosos: variação de porosidade via envelhecimento)
1221Regina Sandra Veiga Nascimento, 1Marcus Vinicius de Araujo Fonseca
1 Universidade Federal do Rio de Janeiro, Cidade Universitária,
Ilha do Fundão, Polo Piloto de Xistoquímica, Instituto de Química,
Centro de Tecnologia, Bl. A, sala 618, Rio de Janeiro, RJ, Brazil, 21949-900.
2 Center for Mineral Technology, Technological Characterization Sector,
Cidade Universitária, Ilha do Fundão, Rua 4, Quadra 3,
Rio de Janeiro, RJ, 21941-590, Brazil
e-mail:perruso@iq.ufrj.br
Keywords: glass leaching, microporous silica
Resumo
Recentemente apresentamos a produção de pós microporosos de sílica de alta pureza a partir de lixiviação ácida de vidros, com tamanho médio de poros próximo de 2,5 nm e 420 m2/g (BET) de área de superfície específica. Foram usados vidros a partir de fusão de dois tipos de resíduos de processamento industrial de xisto oleígeno brasileiro (xisto oleígeno retortado e fração superior da camada intermediária da formação de Irati). Dependendo da proporção dos dois resíduos empregados na formulação dos vidros, é obtida sílica na forma de gel ou de pó, após lixiviação. A lixiviação desses vidros com ácido clorídrico a 90 ºC foi usada para produzir sílica microporosa em pó. Neste trabalho é estudado o efeito do tempo e da temperatura de envelhecimento na morfologia e na estrutura da sílica obtida na forma de pó. Estudos de envelhecimento foram efetuados em dois meios diferentes, um ácido (ácido clorídrico) e um básico (hidróxido de amônio) a diferentes tempos e temperaturas. DRX, MEV/EDX, MET, medidas de área de superfície específica e ATG/ATD foram usados para caracterizar esses materiais. Os resultados mostraram uma diminuição na área de superfície específica com aumento de tempo e de temperatura. Esse comportamento pode ser aparentemente associado a mecanismos de dissolução e re-precipitação.
Palavras-chave: lixiviação de vidros, sílica microporosa
The silica powder production is one of the fundamental raw material industries and has an important role in chemical processing. For some specific applications, finer and less aggregated powders, with higher purity, are needed. For others, microporosity and specific surface area play an important role. Recently, new technologies have been implemented to establish new and improved limits in size, specific surface area and porosity for these materials. The chemical reactions that occur during an aging process change morphology and structure in the silica network, causing strengthening and stiffening. These structural changes that occur during aging have an important effect on the subsequent drying step. The capillary pressure that develops during drying is proportional to the interfacial area of the silica network; if that area is reduced by a process called "coarsening", the maximum pressure generated during drying is smaller. The stronger and stiffer the silica network becomes, the better it can withstand the capillary pressures, so aged silicas crack less. Another important factor comes from the market for supports for catalysts: the specific surface area is very important when this is concerned and the aging process can tailor this property.
It has long been known that the dissolution of nonsilicate constituents may produce a highly porous silica glass [1]. Porous glasses and products manufactured from this material find application in the production of various optical and optoelectronic devices. Other important applications of a special porous glass known as Vycor® may be as virus filters, carriers of chemical catalysts and semipermeable membranes [2]. Crucibles, ultraviolet transmitting tubes, heating bulbs for the cooking of metals etc. can be done by sintering Vycor [2]. Since the appearance of the first patent by Nordberg & Hood [3], for many years all porous glasses produced have been based on alkali borosilicate glasses. Compositions in a certain region of the ternary system (R2O-B2O3-SiO2) give rise, with proper heat treatment, to a two phase system: an acid insoluble (silica rich) and an acid soluble (alkali and boric oxide rich). Out of this region Vycor can not be produced because the interconnected structure can not be properly formed. The stress developed during the leaching process is the main reason for this, giving rise to a collapse of pores whenever the capillary pressures exceed the mechanical strength of the monolithic glass. Thus Vycor is a limited family of borosilicate glasses which may pass satisfactorily through the leaching process even in the form of monoliths, without the appearance of cracks or fractures in the structure after drying.
However, if the goal is not to obtain the intact monolith, more flexible compositions of other glasses may be used, even nonborosilicate ones, especially in the case of powdered glasses which may be used in a subsequent sintering process [4]. Also the drying process may be more flexible.
Therefore, based on the approach of Huggins et al. [5], to compare the chemical durability of glasses by calculating the Si/O ratio, we have chosen, in a previous work [6], the limits at which the glass network can be held together by two adjacent bonds to produce the porous silica glass. At Si/O=0.286 the average number of bridging oxygens per silicon tetrahedron equals to one; and at Si/O=0.333 this number equals to two, meaning that an interconnected polymeric structure may exist.
Based on our previous results of chemical durability for those glasses obtained from oil shale processing wastes [7,8] (Retorted Shale¾RS, and the Top Fraction of Intermediary layer of mining¾TFI) we concluded that it is possible to produce microporous silica powders, with different morphologies, and properties that can vary from the ones presented by the silica powder to the ones shown by the gel, because one of the wastes is silica rich (RS) while the other is alkali rich (TFI). Therefore, through the balance between the amount of network formers and of network modifiers in the glasses composition, it is possible to vary the Si/O ratios. In this study, "frits" of the glasses in which the Si/O ratio bears 0.323 were used to produce the microporous silica powder, and the aging study was performed on the obtained material to evaluate its changes in morphology and structure. Aging was performed in two different media, an acidic (hydrochloric acid) and a basic one (ammonium hydroxide), and two different temperatures, 25 °C and 80 °C, during default font ranging from one to sixteen days. The concentrations of the two media were chosen in order to enhance dissolution/reprecipitation effects [9,10]. A similar study dealing with "frits" of glasses for which the Si/O ratio bears 0.286, resulting in a gelatinous powder, will be published elsewhere.
The chosen formulation was 50%RS/50%TFI by weight, which gives a Si/O ratio of 0.323. The techniques used to characterize the basic structure of the silica powders were: XRD, SEM/EDX, TEM, BET specific surface area measurements and DTA/TGA.
MATERIALS AND METHODS
Si/O ratio calculation. Equation (A), the formula of Huggins et al. [5] for any given composition of glass, was used to calculate the Si/O ratio as follows:
(A)
where SiO2 is the weight % of SiO2; MmOn is the weight % of each oxide, including SiO2; mol MmOn is the molecular weight of each oxide; and On is the number of oxygen atoms in the oxide.
Glass synthesis. The two wastes from oil shale processing (RS and TFI) were individually ball milled and classified by sieving in batches of 5 kg until all the samples had a particle size below 175 mm. After that, equal weights of RS and TFI were simultaneously added to a ball mill (with only a few balls to improve mixing without producing milling) and mixed for 90 minutes. The mixture was melted in ZAS crucibles (batches of ~500 g) at 1450 oC in an electric furnace, holding this temperature for one hour. The melt was poured into a 20 L water container in order to prepare the "frit", which was subsequently dried at 80 oC for 24 hours. After that, the "frit" was ball milled (loads of ~1000 g) and classified by sieving until all the particles had a size below 105 mm.
Leaching. Batches of 100 g were added to 800 mL of 6 N hydrochloric acid (prepared with freshly distilled water), previously stabilized at (90±2) °C. Leaching reactions were conducted for 6 h, as determined by kinetic measurements [8].
Washing. After leaching, the reaction medium was cooled and the powder decanted, shaken with 6 N hydrochloric acid (100 mL) for rinsing and decanted again. The process was repeated once more to eliminate the yellowish coloration of the solution, which is due to the presence of ferric ions. The powder was then shaken with freshly distilled water and decanted. This step was repeated until a pH ~5 was obtained in the solution. Finally the powder was left in water (500 mL) for approximately 20 h, in order to reach the equilibrium inside the microporous cavities.
Drying. The washed powder was decanted and left to dry at room temperature for 48 h. After that, it was transferred to a glass container and left in a oven at 80 oC for 24 h, producing a white powder which was the starting microporous silica used in the aging experiments. This sample was designated S0 and stored in high-density polyethylene containers.
Aging. The experiment was performed in two Pyrex® reactors (500 mL), equipped with sealed reflux systems, one for the acidic medium (300 mL of 2 N hydrochloric acid) and the other for the basic one (300 mL of 2 N ammonium hydroxide). The reactor, also equipped with a sampling window, was kept inside a water bath under controlled temperature (either 25±1°C or 80±1°C). The starting powder, ca. 20 g, was added to the temperature stabilized medium and sampling (ca. 4 g) was performed at default font of 1, 2, 4, 8 and 16 days.
Terminology. In this work the following system of nomenclature for the samples was adopted: A (or AM in some Figures) or B (or BM, ibid.) to indicate, respectively, the acidic or the basic medium in aging; 25C or 80C to indicate the temperature of the bath, 25 °C or 80 °C, respectively; and 1d, 2d, 4d, 8d or 16d to indicate the number of days of aging. For example, A-80C-4d stands for acidic medium at 80 °C during 4 days.
2ndwashing. After aging, the particles were separated by decantation, washed with water (50 mL) and decanted again. The process was repeated until the pH reached neutrality.
2nd drying. Following the 2nd washing, the powder was decanted and left to dry at room temperature for 24 h. After that, it was transferred to a glass container and left in a oven at 80oC for 24 h. The samples were stored in high-density polyethylene containers.
XRD.aplusSEM/EDX. SEM analysis were conducted in a LEICA S440 equipment coupled to an OXFORD LINK ISIS L300 EDX analyzer operated at 20 kV, using a solid state detector with a resolution of 133 eV at 5.9 keV. All samples were sputtered with gold under vacuum.
TEM. TEM analysis were conducted in a JEOL 2000 FX equipment operated at 200 kV. Sample preparation was performed by diluting the samples in deionized water. The support used was a carbon grid with 3 mm of diameter.
BET surface area. Analysis were performed in a Micromeritics ASAP 2010 equipment. Samples were heated to 300 °C under vacuum before measurements.
DTA/TGA. Both analysis were performed in a PERKIN ELMER 7 Series/UNIX equipment, using powdered samples, heated at a rate of 10 oC/min in flowing oxygen atmosphere. The temperature range of study was 25 oC-1200 oC. For both DTA and TGA analysis, samples of 20 mg were used.
RESULTS
According to Equation (A) the calculated Si/O ratio for the 50%RS/ 50%TFI formulation is 0.323. The composition of the 50%RS/50%TFI base glass is shown in the Table I. Note that the humidity present comes from the wet "frits" analyzed (water from the quenching process).
Table 1: Base glass composition. * See text.
The particles from the S0 silica powder are shown in Fig. 1. No great morphological changes can be seen in the fragments of the ball milled glass after leaching. The SEM image of this powder shows particles with smooth surfaces, presenting conchoidal fractures, characteristic of milled (or broken) glass, that remained intact in shape after leaching. Also a small number of cracks on the particles may be seen.
Figure 1: S0 sample showing the starting powder before aging experiments.
SEM images of the particles after the aging processes, regardless of the media, time or temperature used, have shown very similar smooth surfaces, even at high magnifications. However, the TEM of these samples (see the micrograph of B-80C-16d in Fig. 2 for example) clearly show coarsening (or ripening). In Fig. 2 it is shown that the structure of one of the finest particles, selected for TEM image formation (the scale bar represents 50 nm), consists of a three dimensional network. The channels that were formed in the leaching process gave rise to this microporous, silica rich, network. The arrows indicate rounded necks enlarged after aging.
Figure 2: B-80C-16d sample showing coarsening. Arrows indicate rounded necks. Scale bar represents 50nm.
Semiquantitative EDX spectra (Fig. 3) show the presence, in the base glass, of all the elements which were previously mentioned in Table I (in the form of oxides). It is also shown the evolution to a silica rich powder, with no other elements detected (down to the detection limit of ~0.1% by wt.). The same result was obtained for all the aged powders in EDX experiments.
Semiquantitative EDX spectra showing the elements present in the base glass and in the S0 sample (Au due to sputtering).
In acidic medium, a slight increase in specific surface area could be observed after 1 day of aging, this being slightly higher at 80 °C than at 25 °C. After 16 days, the surface area decreased slightly at 25 °C, but was reduced to ~44% of the starting materials surface area when aged at 80 °C.
Fig. 4 shows the XRD pattern of the starting powder S0, the base glass, and a standard of quartz. In S0 and in all aged powders there is an increase of counts in the region typical of amorphous materials. A high background in all patterns is also associated with the presence of amorphous materials. However, the base glass and the S0 powder are not completely amorphous. There is a peak which corresponds to quartz in the XRD pattern (see Fig. 4 and the region in detail). The same was observed in the aged samples, with no significant differences.
Specific surface area measurements (BET, three points) on S0 reference sample and on the ones with 1 day and with 16 days of aging were performed for each series of samples. The results may be seen in Fig. 5. As a general rule, the BET specific surface area of samples decreased with aging time.
BET results of some aged samples (BM and AM meaning the basic and the acid media of aging).
Aging in basic media, at 25 °C, produced a decrease to ~87% after 1 day and to 75% after 16 days. At 80 °C, however, the surface area increased slightly after 1 day of aging, but dropped to ~25% of the starting materials surface area after 16 days.
Some DTA results are shown in Fig. 6 presenting the general behavior of the aged samples and of S0. Note that curves were displaced intentionally from each other by 1 °C for clarity. No well defined peaks were observed in the region below 1200 °C, except that related to free water desorption, whose related peak maxima were in the range ~100 °C to ~130 °C.
Also, we see no evidence that could be associated to the slow water release process which takes place at temperatures above ~150 °C. That process occurred slowly over the temperature range studied (mainly below ~800 °C). It is interesting to point out that no peak of crystallization was found for the aged samples .
TGA (and DTG) of all samples showed the same pattern, revealing two distinct regions of mass loss: below and over 150 oC. The total water released from samples has varied over a range of ~10% to ~15% as a result of the previous drying process (at 80 °C) which is not so effective and reproducible.
The amount of water released below 120 °C-150 °C is related to adsorbed water molecules. The amount released above 150 °C is shown as a function of time, temperature and medium of aging in Fig. 7. This water derives from the silanol surface groups of the aged silica powders, directly related to the specific surface area. The point on the y-axis corresponds to the starting powder S0.
The mass loss of water decreased with the specific surface area (see BET results). The results show a decrease in mass loss as time of aging increased. In the samples aged at 80 °C the temperature effect was more pronounced after the longest period of time (16 days), indicating a significant decrease of water content in relation to the 25 °C aged samples. The surface structure of S0 seems to be not comparable to the structure of the aged powders.
DISCUSSION
According to Huggins et al. [5] the Si/O ratio may be broadly considered as the measure of the polymerization degree of the network. The calculated value for the glass under study suggests that it would have poor resistance to chemical agents. The resultant microporous powder found after the leaching process, the S0 silica rich powder, confirmed that supposition [6].
From our previous studies of this glass in the monolithic form [6-8] the time selected for the leaching process was sufficient to completely remove the nonsilicate constituents. This could be confirmed by the semiquantitative EDX microanalysis results, which showed, for the S0 sample and the aged ones, that no elements other than silicon and oxygen were present down to the detection limit of ~0.1% by weight. Indirectly, the purity of the samples may be correlated to the DTA results that did not show any peak of crystallization of any phase of silica, which would easily occur if some network modifiers were present, even in trace amounts [9]. The results indicated that the quartz phase seems to be in an early stage of formation, since only the quartz related peak of highest intensity could be observed.
The leaching process itself seems to improve the amorphization of the overall structure resultant of the obtained glass, as evidenced by the increase of counts in the region of the XRD spectrum related to amorphous materials. However this process did not show any evolution with time, temperature or media.
The aging process did not produce any macro morphological changes as observed in SEM images of all the aged samples. However, the TEM images showed different stages of the coarsening process of the silica skeleton of the powders.
Coarsening is a process of dissolution and reprecipitation driven by differences in solubility between surfaces with different radii of curvature [10]. As the solubility and rate of dissolution of silica increases rapidly at high pH, it is not surprising that the rate of coarsening is similarly pH-dependent [9]. The effect of temperature also contributes in the same direction as the solubility of silica increases with increasing temperatures [9].
Particles have positive radii of curvature and hence they are more soluble than a flat plate of the same material. Crevices and necks between particles have negative radii, so their solubility is specially low and material accumulates at these sites. The result of the dissolution-reprecipitation process is the reduction of the net curvature of the solid phase and therefore a decrease of the interfacial energy and surface area. As a consequence, the necks become round shaped and larger (coarser) and the surfaces smoother. The more evident case is for the sample B-80C-16d seen in Fig. 2. For this sample, it was employed the highest temperature and pH and the longest aging time, resulting in the greatest reduction of the specific surface area, relative to the starting silica powder S0. Associated with it, very little water was released over ~150 °C (only comparable to the A-80C-16d sample).
BET measurements showed that the specific surface area of samples decreased with aging time (comparing the results for 1 day with the ones for 16 days of aging, in each series of experiments). This is in good agreement with the coarsening theory. For the experiments at 80 °C, only the samples submitted to 16 days of aging exhibited lower specific surface areas than the ones treated at 25 °C. This also is in good agreement with the coarsening theory. However, at 1 day of aging in the acidic medium, the specific surface area of the samples did not show significant difference. Since the solubility of the silica is very low [9] it is not surprising that significant changes in aging were noticed only after 16 days. In the basic medium, the sample aged at 80 °C for 1 day (B-80C-1d) showed an increase of specific surface area relative to S0 and to B-25C-1d. Since the solubility of silica at high pH and temperature is very high, an accentuated dissolution process may explain this result i.e., dissolution takes place until the solubility of the silica in the medium is reached. It seems that the necessary equilibrium condition was not attained for the occurrence of a significant reprecipitation step in the coarsening mechanism (a dynamic equilibrium). Then a condition to an increase in specific surface area is created in the beginning of the aging process.
This explanation is in accordance with the TGA data which shows an increase in mass loss (water resulting from surface hydroxyl removal) when the aging time was increased from one to two days. This process seems to be associated with temperature (the same behavior was observed with the acidic medium at 80 °C) probably caused by the increase in solubility, which then demands more time to reach the equilibrium (however this is faster under basic medium that in acidic medium, but the difference could not be seen in the time scale selected). After the equilibrium is reached the result is a decrease in mass loss in accordance with the coarsening theory.
CONCLUSIONS
The Si/O ratio calculation proved to be useful on the prediction of the low chemical resisting behavior of the studied glasses.
Semiquantitative EDX measurements together with the DTA technique could show that the time selected for the leaching process was sufficient to completely remove the nonsilicate constituents giving rise to a microporous silica powder.
Aging experiments have been performed on a microporous silica powder and the coarsening theory seems appropriate to explain all the results of the TEM, BET and TGA techniques, which have shown a good correlation between them.
As a general behavior, an increase in time and temperature of the aging process causes a reduction of the specific surface area of the silica powder samples. A non-equilibrium state could be promoting an initial increase of the specific surface area of the samples aged at 80 °C.
ACKNOWLEDGMENTS
The authors thank Viviane Menezes (IQ-UFRJ) for DTA/TGA, Marcia Soares Sader for TEM (COPPE-UFRJ) and Marcelo Correa Andrade (LQS-CETEM) for BET measurements.
(Rec. 02/98, Rev. 05/98, Ac. 05/98)
- [1] W. E. S. Turner, F. Winks, The influence of boric oxide on properties of chemical and heat-resisting glasses. II. The resistance to chemical reagents. J. Soc. Glass Tech. 10, (1926)102-113.
- [2] M. B. Volf, Technical Glasses. Prague, SNTL, 1961
- [3] M. E. Nordberg, H. P. Hood, US Patent 2,106,744 ( Feb. 1, 1938).
- [4] J. Y. Ying, J. B. Benziger, A. Navrotsky, Structural evolution of colloidal silica gels to glass. J. Am. Ceram. Soc., 76, 10, (1993) 2561-70.
- [5] M. L. Huggins, K-H. Sun, A. Silverman, The vitreous state. J. Am. Ceram. Soc., 26, 12 (1943) 393-8.
- [6] C. R. Perruso, R. Neuman, M. V. A. Fonseca, Synthesis and structural characterization of microporous glasses. In: 41o Congresso Brasileiro de Cerâmica, June 3-6, 1997, Săo Paulo, in press.
- [7] C. R. Perruso, M.V. A. Fonseca, Primeiro Simpósio Nacional de Vidro,1 (1995)103-4.
- [8] C. R. Perruso, D. Sc. Thesis, in progress.
- [9] R. K. Iler, The chemistry of silica, New York, N.Y.: Wiley, 1979.
- [10] C. J. Brinker, G. W. Scherer, Sol-gel science: the physics and chemistry of sol-gel processing. San Diego, CA.: Academic Press, 1990.
Publication Dates
-
Publication in this collection
12 June 2000 -
Date of issue
Oct 1998
History
-
Accepted
May 1998 -
Received
Feb 1998 -
Reviewed
May 1998
